In the spring of last year, we introduced Anthropic's AI for Science program, aiming to enhance scientific research and innovation through our API offerings. Since its inception, this initiative has aided researchers on numerous impactful projects, including drug repurposing and quantum simulation. Our observations indicate that collaboration among multiple grantees addressing similar research questions leads to more productive outcomes. To facilitate this, we are initiating thematic calls for proposals within the overarching AI for Science framework.
Today, we are excited to announce a targeted call for proposals focused on rare genetic diseases. Successful applicants will be awarded up to $50,000 in Claude credits over a six-month period, with the intention of fostering a community among researchers dedicated to exploring how AI can transform our understanding of these rare conditions. This program will offer two distinct avenues: one for researchers engaged in foundational science and another for early-stage biotech companies focused on expediting clinical development for rare diseases.
Research on rare diseases is particularly challenging due to the limited knowledge surrounding basic scientific principles. Collectively, rare diseases affect a significant portion of the global population—approximately 400 million individuals are living with one of over 7,000 distinct rare diseases. However, the dispersed nature of these conditions complicates clinicians' efforts to establish patient registries, identify potential therapeutic targets, and design effective clinical trials. Additionally, the unique characteristics of rare diseases, such as specific genetic variations or symptom combinations, often lead to isolated studies that make it difficult to identify common mechanisms across different diseases. As with all drug development, rare diseases also encounter the considerable hurdle of lengthy timelines required to transition promising drug candidates into patient trials.
We believe AI can provide valuable assistance in addressing these challenges. With its capabilities to model rare genetic conditions accurately and identify patterns, AI enables researchers to synthesize findings from vast amounts of literature, rapidly extract insights from small datasets, and standardize terminology. These advancements are crucial as researchers strive to maximize the utility of existing information while simultaneously working to generate new data, navigate access issues, and manage regional disparities in research efforts. Therefore, we are focusing our current call for AI for Science projects specifically on rare diseases.
**Track One: Advancing Basic Science Collaborations**
The first area of our rare disease research funding program is designed to promote partnerships among clinical researchers, patient organizations, and data scientists, ultimately accelerating the understanding of the mechanisms behind rare diseases.
A key collaborator in this initiative is the Monarch Initiative, a global consortium dedicated to enhancing the diagnosis and identification of underlying mechanisms for patients with rare disorders. Monarch has developed vital resources like the Mondo Disease Ontology, a coding system that harmonizes disease definitions from various sources, and the Monarch Knowledge Graph, which consolidates genotype-phenotype relationships across species to assist with diagnostics and mechanism exploration.
Recently, contributors to Monarch have begun creating DisMech, an agent-friendly database that allows Claude to analyze case reports, variant databases, and public data more rapidly than ever, highlighting mechanistic similarities between diseases. Monarch is inviting our AI for Science grantees to utilize and enhance resources like Mondo and DisMech to propose new mechanistic hypotheses that could aid in treatment development.
While Monarch's efforts to augment the interoperability of rare disease information can benefit significantly from Claude’s capabilities, we recognize that there is still much work to be done in terms of data collection, enhancing diagnostic infrastructure, and fostering patient-led initiatives throughout the rare disease ecosystem. We remain committed to collaborating with Monarch and other organizations to tackle challenges where the application of AI is less straightforward, sharing our findings along the way.
**Track Two: Supporting Biotechnology Innovations**
The second focus of our rare disease research funding program is to assist biotechnology experts and emerging companies aiming to fast-track drug development for rare diseases. Currently, it can take one to two years to transition from a confirmed genetic diagnosis to an available treatment, largely due to delays in securing manufacturing capabilities, sequential running of safety studies, and assembling the extensive regulatory documentation needed for patient trials.
We believe Claude has the potential to significantly accelerate certain phases of this process, especially in streamlining documentation, speeding up therapeutic strategy selection by assessing the druggability of targets across various modalities, and identifying shared mechanisms among genetic therapies, potentially enabling "basket trials" that would eliminate the need for separate IND applications for each patient.
Although many facets of drug development cannot be expedited due to manufacturing or safety testing constraints, we are optimistic that considerable progress can be achieved. By providing API credits and access to Claude Science for numerous biotechnologists and startups, we aim to inspire innovative explorations and solutions.
We also anticipate that our program’s grantees will build upon the work of existing partners in rare disease therapeutics. For instance, Every Cure, one of our current AI for Science grantees, utilizes Claude to discover drug repurposing possibilities across millions of candidates. Likewise, the Centre for Population Genomics—a partnership between the Garvan Institute and the Murdoch Children’s Research Institute—is developing a Claude-based system for drafting variant classifications, addressing one of the significant bottlenecks in diagnosing rare genetic conditions. Additionally, the Violet Research Institute, a small nonprofit dedicated to researching ultra-rare genetic disorders, relies on Claude for navigating FDA guidelines, conducting bioinformatics analyses, and creating regulatory filings.
**Next Steps**
To apply for funding in either track of the AI for Science rare disease research program, interested parties should complete this application. We will accept applications until August 2, 2026, at 11:59 PM PST. Selected applicants can utilize their credits to access Claude Opus or other models approved for biological applications. Projects facing challenges with our bio classifiers might qualify for exemptions.
Examples of potential projects in Track One include proposing mechanisms linking distinct rare diseases with shared genes or pathways, curating patient organization data to enhance natural history studies, and evaluating models for effectiveness in rare disease tasks.
For Track Two, project examples might include justifying starting doses based on limited data, mining natural history data to pinpoint measurable biomarkers, and drafting regulatory documentation to expedite the review process.
This rare disease grant program aligns directly with our mission to leverage AI for beneficial outcomes in areas that may not naturally receive focus through market dynamics. Nonetheless, tackling the complexity of rare diseases requires collaborative effort; neither a single organization nor one approach can address it entirely. While Claude can help expedite therapeutic timelines and streamline biological data curation, there are limitations when data is scarce or poorly formatted, as well as challenges related to insurance authorization and access to diagnostic facilities. We look forward to seeing how our new grantees collaborate with one another and our existing AI for Science partners to push the boundaries of basic science and contribute meaningfully to broader scientific endeavors.
